RESEARCH Open Access The clinical usefulness of ...

RESEARCH Open Access

The clinical usefulness of extravascular lung waterand pulmonary vascular permeability index todiagnose and characterize pulmonary edema: aprospective multicenter study on the quantitativedifferential diagnostic definition for acute lunginjury/acute respiratory distress syndromeShigeki Kushimoto1*, Yasuhiko Taira2, Yasuhide Kitazawa3, Kazuo Okuchi4, Teruo Sakamoto5, Hiroyasu Ishikura6,Tomoyuki Endo7, Satoshi Yamanouchi1, Takashi Tagami8,9, Junko Yamaguchi10, Kazuhide Yoshikawa11,Manabu Sugita12, Yoichi Kase13, Takashi Kanemura14, Hiroyuki Takahashi15, Yuichi Kuroki16, Hiroo Izumino17,Hiroshi Rinka18, Ryutarou Seo19, Makoto Takatori20, Tadashi Kaneko21, Toshiaki Nakamura22, Takayuki Irahara23,Nobuyuki Saito24 and Akihiro Watanabe8, for The PiCCO Pulmonary Edema Study Group

Abstract

Introduction: Acute lung injury (ALI)/acute respiratory distress syndrome (ARDS) is characterized by features otherthan increased pulmonary vascular permeability. Pulmonary vascular permeability combined with increasedextravascular lung water content has been considered a quantitative diagnostic criterion of ALI/ARDS. Thisprospective, multi-institutional, observational study aimed to clarify the clinical pathophysiological features ofALI/ARDS and establish its quantitative diagnostic criteria.

Methods: The extravascular lung water index (EVLWI) and the pulmonary vascular permeability index (PVPI) weremeasured using the transpulmonary thermodilution method in 266 patients with PaO2/FiO2 ratio ≤ 300 mmHg andbilateral infiltration on chest radiography, in 23 ICUs of academic tertiary referral hospitals. Pulmonary edema wasdefined as EVLWI ≥ 10 ml/kg. Three experts retrospectively determined the pathophysiological features ofrespiratory insufficiency by considering the patients’ history, clinical presentation, chest computed tomography andradiography, echocardiography, EVLWI and brain natriuretic peptide level, and the time course of all precedingfindings under systemic and respiratory therapy.

Results: Patients were divided into the following three categories on the basis of the pathophysiologicaldiagnostic differentiation of respiratory insufficiency: ALI/ARDS, cardiogenic edema, and pleural effusion withatelectasis, which were noted in 207 patients, 26 patients, and 33 patients, respectively. EVLWI was greater in ALI/ARDS and cardiogenic edema patients than in patients with pleural effusion with atelectasis (18.5 ± 6.8, 14.4 ± 4.0,and 8.3 ± 2.1, respectively; P < 0.01). PVPI was higher in ALI/ARDS patients than in cardiogenic edema or pleuraleffusion with atelectasis patients (3.2 ± 1.4, 2.0 ± 0.8, and 1.6 ± 0.5; P < 0.01). In ALI/ARDS patients, EVLWI increasedwith increasing pulmonary vascular permeability (r = 0.729, P < 0.01) and was weakly correlated with intrathoracicblood volume (r = 0.236, P < 0.01). EVLWI was weakly correlated with the PaO2/FiO2 ratio in the ALI/ARDS and

* Correspondence: [email protected] of Emergency Medicine, Tohoku University Graduate School ofMedicine, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, JapanFull list of author information is available at the end of the article

Kushimoto et al. Critical Care 2012, 16:R232http://ccforum.com/content/16/6/R232

© 2013 Kushimoto et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.

mailto:[email protected]

http://creativecommons.org/licenses/by/2.0

cardiogenic edema patients. A PVPI value of 2.6 to 2.85 provided a definitive diagnosis of ALI/ARDS (specificity, 0.90to 0.95), and a value < 1.7 ruled out an ALI/ARDS diagnosis (specificity, 0.95).

Conclusion: PVPI may be a useful quantitative diagnostic tool for ARDS in patients with hypoxemic respiratoryfailure and radiographic infiltrates.

Trial registration: UMIN-CTR ID UMIN000003627

IntroductionPulmonary edema is characterized by the abnormal accu-mulation of fluid in the extravascular space of the lungsand is a common finding in critically ill patients [1]. Thispathological condition may develop due to an increase inthe pulmonary capillary permeability (acute lung injury(ALI), acute respiratory distress syndrome (ARDS)), anincrease in the pulmonary capillary hydrostatic pressure(hydrostatic or cardiogenic pulmonary edema), or both[2]. Pulmonary edema can be detected by clinical evalua-tion of factors such as patients’ history, physical findings,and routine laboratory examinations, and is confirmed bythe presence of bilateral pulmonary infiltration on chestradiography [2,3]. However, interpretation of these factorsis often limited by a certain degree of subjectivity thatmight cause inter-observer variation even among experts,particularly in critically ill patients [4-6]. Moreover, inten-sive care physicians may find it difficult to determine thecause of the extravascular lung water (EVLW) increase [7].In 1994 the American Thoracic Society and the Eur-

opean Society of Intensive Care Medicine co-publishedthe proceedings of a consensus conference on ARDS, anddefined ALI and ARDS as an American-European Consen-sus Conference (AECC) definition [8,9]. Although manyclinical trials performed after the publication of the pro-ceedings used the AECC definition, this definition hasbeen suggested to have various issues, including a lack ofexplicit criteria for defining what is acute, the sensitivity ofthe PaO2/FiO2 (P/F) ratio to different ventilator settings,poor reliability of the chest radiograph criterion, and diffi-culties distinguishing hydrostatic edema [10-14]. Thesecriteria are also not sensitive predictors of disease severityand outcomes [4,5,15-17] because: the P/F ratio variesconsiderably across different FiO2 levels, particularly whenFiO2 < 0.5, PaO2 > 100 mmHg, or when the shunt fractionis low; many patients who initially fulfill the ARDS criteriamight improve the P/F ratio > 200 mmHg after applicationof positive end-expiratory pressure for a short time or theuse of higher FiO2; and hypoxemia in ARDS may also berelated to atelectasis or a low cardiac output [14]. Basedon these limitations, a novel definition has been proposedthat takes into account the clinical and physiologic charac-teristics of ALI/ARDS [18]. The Berlin definition forARDS was published recently and was demonstrated tohave better predictive validity for mortality than the AECC

definition [10]. Although ARDS has been described as atype of acute, diffuse inflammatory lung injury leading toincreased pulmonary vascular permeability, increased lungweight, and loss of aerated lung tissue, as the panel agreedin their conceptual model, none of the suggested criteriaevaluates the increase in pulmonary microvascular perme-ability, a hallmark of ARDS [10]. Not only the AECC defi-nition but also the Berlin definition may include anextensive range of respiratory insufficiencies without anincrease in pulmonary microvascular permeability.Previous studies have reported on methods of quanti-

fying pulmonary edema [19,20]. The double-indicatorthermodilution technique allows the measurement ofEVLW, and excellent correlation between in vivo andpostmortem gravimetric EVLW values was obtained inboth animal and human studies using this method[21,22]. However, this method is cumbersome and tech-nically challenging for routine clinical application. Thesingle-indicator technique is therefore used in clinicalsettings; this method is as sensitive as the double-indica-tor technique [23,24]. We previously validated theaccuracy of EVLW measurements obtained using thesingle-indicator technique in the postmortem lung sam-ples and defined the statistically normal EVLW values ina human autopsy study [25]. The close relationshipbetween EVLW and outcome has been also demon-strated [26].The transpulmonary thermodilution technique pro-

vides an estimation of both EVLW and the pulmonaryblood volume, and the ratio of these two parameters -the pulmonary vascular permeability index (PVPI) - hasbeen shown to reflect the pulmonary microvascular per-meability [7,27].Increased pulmonary vascular permeability is the cru-

cial pathophysiological feature of ALI/ARDS and hasbeen considered a quantitative diagnostic criterion forALI/ARDS [28]. PVPI has been evaluated to enable oneto differentiate ALI/ARDS from hydrostatic edema [7].PVPI was shown to be useful for determining themechanism of pulmonary edema in ALI/ARDS, andPVPI ≥ 3 allowed the diagnosis of ALI/ARDS with asensitivity of 85% and a specificity of 100%. However,that study was a single-center retrospective review ofonly 48 patients (ALI/ARDS, 36 patients; hydrostaticedema, 12 patients).


of 15

The aims of this study were to clarify the clinicalpathophysiological features of ALI/ARDS, and to estab-lish the quantitative differential criteria of ALI/ARDS onthe basis of PVPI assessed using the transpulmonarysingle thermodilution technique.

Materials and methodsThis prospective, observational, multi-institutional studywas approved by the ethics committee of each of the 23institutions, and written informed consent was providedby all patients’ next of kin. The study was registeredwith the University Hospital Medical Information Net-work Clinical Trials Registry: UMIN-CTR IDUMIN000003627.

PatientsBetween March 2009 and August 2011, 301 patientsfrom 23 critical care centers at tertiary care hospitalswere enrolled in this study. In all of 23 participatinginstitutions, the single transpulmonary thermodilutiontechnique is one of the standard monitoring methodsfor circulatory and respiratory management of criticallyill patients. The median (interquartile range) number ofincluded patients per each institution was 10 (6 to 18).The inclusion criteria were aged older than 15 years,

requiring mechanical ventilation (expected over 48 hours)for acute respiratory failure with a P/F ratio ≤ 300 mmHgand bilateral infiltration on chest radiography and transpul-monary thermodilution technique monitoring of circula-tory/respiratory status as per the attending physician’sdiscretion. Exclusion criteria were as follows: over 5 days

from the onset of acute respiratory failure with a P/F ratio≤ 300; chronic respiratory insufficiency (chronic obstruc-tive pulmonary disease, and so forth); history of pulmonaryresection/pneumonectomy, pulmonary thromboembolism,and severe peripheral arterial disease; cardiogenic shockwith a cardiac index < 1.5 l/minute/m2; acute phase oftrauma with lung contusion and burns; other causes ren-dering patients unsuitable for evaluation with the transpul-monary thermodilution technique, including patients withmoderate to severe valvopathy; and the attending physicianidentifying patients as inappropriate for inclusion.Of the 301 enrolled patients, 266 were included in this

analysis. Reasons for the exclusion of 35 patients areshown in Figure 1.In this study, the diagnosis of pulmonary edema was

established on the basis of: the presence of bilateralinfiltrates on chest radiography; P/F ratio ≤ 300 mmHg;and an increase in the EVLW indexed to the predictedbody weight (extravascular lung water index (EVLWI))≥ 10 ml/kg. Although there is no definitive quantitativecriterion of EVLWI for pulmonary edema, we recentlyreported that the normal EVLWI value is approximately7.4 ± 3.3 ml/kg for humans [25]. EVLWI ≥ 10 ml/kgwas used for definition of pulmonary edema in the pre-viously reported study [1,29].

Measurement of EVLWI and PVPI using thetranspulmonary thermodilution methodA 4-Fr or 5-Fr femoral arterial or 4-Fr brachial arterialthermistor-tipped catheter (PV2014L16N, PV2015L20N,or PV2014L22N; Pulsion Medical Systems, Munich,

301 patients enrolled

207 ALI/ARDS

26 cardiogenic edema

33 pleural effusion with atelectasis

35 excluded14, respiratory insufficiency secondary to sepsis without elevation of EVLWi

to >10 mL/kg due to hypovolemia8, unclassified by expert decision5, atelectasis and consolidation due to pneumonia3, acute phase of trauma with lung contusion/burn2, nonacute onset1, near drowning1, combined with COPD1, after lung resection

Figure 1 Patient enrollment, exclusion, and classification. ALI, acute lung injury; ARDS, acute respiratory distress syndrome; COPD, chronicobstructive pulmonary disorder; EVLWI, extravascular lung water index.


of 15

Germany) was inserted in all patients by the attendingphysicians’ discretion and connected to the PiCCO® plusor PiCCO® 2 monitor (Pulsion Medical Systems). Themonitor uses a single-thermal indicator technique to cal-culate the cardiac output (CO), global end-diastolicvolume (GEDV), EVLW, and other volumetric parameters.A 15 ml bolus of 5% glucose at 5°C was injected through acentral venous catheter, and the CO calculated using theStewart-Hamilton method. A 15 ml bolus dose wasselected as previously described [25] and the precision ofmeasurements has been demonstrated [30,31]. The centralvenous catheters were accessed from the jugular or subcla-vian route in all patients. Concurrently, the mean transittime and exponential downslope time of the transpulmon-ary thermodilution curve were calculated. The product ofCO and mean transit time represents the intrathoracicthermal volume [23]. The product of CO and exponentialdownslope time represents the pulmonary thermal volume[32]. GEDV is calculated as the difference between theintrathoracic thermal volume and the pulmonary thermalvolume, and represents the combined end-diastolicvolumes of the four cardiac chambers. The intrathoracicblood volume (ITBV) is calculated as the linear relation-ship with the GEDV [23]:

ITBV = 1.25 × GEDV − 28.4

EVLW is the difference between the intrathoracicthermal volume and ITBV [23,32]. The detailed princi-ples and calculations involved in deriving EVLW usingthe thermodilution technique are discussed elsewhere[7,33]. PVPI is calculated as the ratio of EVLW and pul-monary blood volume [7]. ITBV and GEDV are indexedto the body surface area.The median EVLW value was obtained after three bolus

injections of 15 ml each [31]. The absolute EVLW valuewas indexed to predicted body weight, calculated as 50 +0.91 (height (cm) - 152.4) for males and 45.5 + 0.91 (height(cm) - 152.5) for females [34]. For indexing EVLW, thepredicted body weight instead of the actual body weightwas used because: lung volumes are dependent on genderand height, not on weight [35]; measurement of EVLWindexed to the actual body weight can be underestimatedin obese patients [36,37]; and the EVLWI has been shownto be a better prognostic indicator than EVLW indexed tothe actual body weight [38-40]. The results were analyzedusing PiCCO-VoLEF Data Acquisition for Win32 Version6.0 for PiCCO® plus or Version 2.0.0.13 for PiCCO®

2 (Pulsion Medical Systems).

Assessment of circulatory/respiratory status, otherparameters, and clinical courseAt the time of enrollment (day 0), the clinical conditions,cause of respiratory insufficiency, Acute Physiology andChronic Health Evaluation II score, Sequential Organ

Failure Assessment score, and Lung Injury Scale scorewere assessed [41,42]. Echocardiography was performed tomeasure the left ventricular ejection fraction, left ventricu-lar end-diastolic dimension, interventricular septum thick-ness, E/A ratio, left atrial dimension, inferior vena cavadiameter and its respiratory variation, presence of hypo/akinesis, valvular abnormality, left ventricular systolic/dia-stolic function, and the thermodilution hemodynamicassessment validity. Chest computed tomography was alsoconducted on the day of enrollment. B-type natriureticpeptide (BNP) or N-terminal pro-BNP was measured onthe day of enrollment and daily thereafter.From the day of enrollment to day 2, the circulatory/

respiratory status and other parameters except for rou-tine clinical workup were assessed; the clinical course,including respirator settings, Lung Injury Scale score[43], Sequential Organ Failure Assessment score, antith-rombin activity level, serum procalcitonin level, dailyfluid intake/output and balance, and therapeutic inter-ventions (surgery, antibiotics, steroids, diuretics, renalreplacement therapy, and so forth) were recorded daily.All patients were followed for 28 days after enrollment

and assessed for 28-day all-cause mortality.

Determination of pathophysiological diagnosticdifferential of respiratory insufficiencyAt least three experts (intensive care, respiratology, andcardiology) retrospectively determined the pathophysiolo-gical mechanism of respiratory insufficiency: ALI/ARDS,increased pulmonary vascular permeability without orwith increased pulmonary vascular hydrostatic pressure;cardiogenic edema, increased pulmonary capillary hydro-static pressure without increased vascular permeability;and pleural effusion with atelectasis, no evidence of lungedema secondary to increased hydrostatic pressure or vas-cular permeability, as previously reported [7,44]. For thispurpose, the experts particularly considered the patients’medical history, clinical presentation and the course, chestcomputed tomography and radiography findings, echocar-diography findings, and serum BNP or N-terminal pro-BNP and procalcitonin concentrations, and systemicinflammatory status. The pulmonary capillary wedge pres-sure was not routinely obtained in this study and anypressures measured depended on the attending physicians’discretion and was only obtained for selected patients. Thephysicians also considered in particular the time course ofall the preceding findings, including daily fluid intake/out-put and balance, hemodynamic parameters obtained fromthe thermodilution method, the requirement for systemicmanagement and respiratory therapy, and the clinicalresponses to these treatments. In this study, EVLWI≥ 10 ml/kg was used for definition of pulmonary edema,and patients with EVLWI ≥ 10 ml/kg were discriminatedbetween permeability and hydrostatic pulmonary edema.


of 15

The experts who provided the final diagnosis were com-pletely blinded to the PVPI, but not to hemodynamic para-meters such as cardiac index, stroke volume index, ITBVand EVLWI.

Statistical analysisData are presented as mean ± standard error or as med-ian (interquartile range) depending on the distributionnormality of the variables. Spearman’s rank correlationwas used for determining the correlation between twovariables, and Mann-Whitney’s U test was used forassessing the differences between two groups. For multi-ple-group comparison, analysis of variance on rankswith a Tukey honestly significant difference test wasused. Receiver operating characteristic curves were gen-erated for PVPI and ITBV by varying the discriminatingthreshold of each parameter. The area under the recei-ver operating characteristic curve for each parameterwas calculated and compared using a Hanley-McNeiltest. P < 0.05 was considered significant. All statisticalanalyses were performed using SPSS 19.0 for Windows(SPSS, Chicago, IL, USA).

ResultsPatient characteristicsThe most frequent condition leading to exclusion ofpatients was respiratory insufficiency (P/F ratio ≤ 300mmHg and slight bilateral infiltration) secondary to sep-sis suggesting ALI but not accompanied by EVLWI ≥ 10ml/kg - the predefined value for pulmonary edema, dueto hypovolemia. Consensus on inclusion of such patientswas not obtained from all attending experts.For this analysis, 266 patients were included and

divided into the following three categories on the basisof the pathophysiological diagnostic differentiation ofthe respiratory insufficiency: ALI/ARDS, cardiogenicedema (including fluid overload), and pleural effusionwith atelectasis. ALI/ARDS complicated by increasedhydrostatic pressure was judged as ALI/ARDS.Table 1 presents the patient characteristics at the time of

enrollment. No patient with body mass index ≥ 30 wasincluded. The incidence of sepsis as a baseline conditionwas higher in the ALI/ARDS patients than in cardiogenicedema and pleural effusion with atelectasis patients. Theperiod of ventilator-free days within 28 days was signifi-cantly longer in patients with cardiogenic edema. Mortal-ity was assessed by 28-day all-cause mortality and therewas no significant difference between the three groups.On the day of enrollment, ALI/ARDS patients had higherAcute Physiology and Chronic Health Evaluation II andSequential Organ Failure Assessment scores than patientswith cardiogenic edema, and had higher Lung Injury Scalescores than patients with pleural effusion with atelectasis.ITBV was highest in the cardiogenic edema patients.

Table 2 shows the underlying condition and mechanismin patients with ALI/ARDS, with 128 of 207 cases causedby sepsis. The most frequent site of infection was therespiratory tract, and in 125 patients the injury was direct.

Comparison of extravascular lung water index andpulmonary vascular permeability indexThe EVLWI on the day of enrollment was significantlyhigher in ALI/ARDS patients than in patients with pleuraleffusion with atelectasis (18.5 ± 6.8 vs. 8.3 ± 2.1; P < 0.01)or cardiogenic edema (14.4 ± 4.0; P < 0.01) (Figure 2). ThePVPI on the day of enrollment was higher in the ALI/ARDS patients than in cardiogenic edema or pleural effu-sion with atelectasis patients (3.2 ± 1.4, 2.0 ± 0.8, and1.6 ± 0.5, respectively). Although the EVLWI was higherin the cardiogenic edema than in pleural effusion withatelectasis patient (Figure 2), there was no significant dif-ference in PVPI between those groups (Figure 3).These differences were also noted when the maximal

values of EVLWI and PVPI recorded during the studyperiod were compared among the three groups (Figures2 and 3).

Relationship among EVLWI, PVPI, and intrathoracic bloodvolumeFor this analysis, cardiogenic edema and pleural effusionwith atelectasis patients were considered non-ALI/ARDSpatients because increased pulmonary vascular perme-ability is not the pathogenetic mechanism of these con-ditions and was not elevated compared with that inALI/ARDS patients.In the ALI/ARDS patients, a strong correlation between

EVLWI and PVPI (r = 0.729, P < 0.01) and a weak correla-tion between EVLWI and ITBV (r = 0.236, P < 0.01) werenoted on the day of enrollment (Figure 4). In the non-ALI/ARDS patients, moderate correlations between EVLWIand PVPI (r = 0.464, P < 0.01) and between EVLWI andITBV (r = 0.493, P < 0.01) were noted (Figure 5).Multiple regression analysis was also performed using

EVLWI as the dependent variable with PVPI and ITBVas the independent variables. The standardized partialregression coefficients were 0.958 for PVPI and 0.646for ITBV in ALI/ARDS patients, and were 0.836 and0.814 in non-ALI/ARDS patients, respectively, suggest-ing the important contribution of PVPI on the EVLWIin ALI/ARDS.

Relationship between extravascular lung water index andPaO2/FiO2 ratioFor this analysis, patients with pleural effusion withatelectasis were excluded because the increased EVLWis not the pathogenetic mechanism of this condition,and EVLWI in these patients was not high as in thosepatients with ALI/ARDS and cardiogenic edema.


of 15

The P/F ratio varied widely at all levels of EVLWI inpatients with both ALI/ARDS and cardiogenic edema(Figure 6). A negative but weak correlation was notedbetween EVLWI and the P/F ratio in all patients exceptfor those with pleural effusion with atelectasis (r =-0.213, P < 0.01) and ALI/ARDS (r = -0.215, P < 0.01).No correlation was found between EVLWI and PVPI incardiogenic edema patients (r = -0.176, P = 0.39).

Differential diagnosis of ALI/ARDS on the basis ofpulmonary vascular permeabilityReceiver operating characteristic curves were generatedusing PVPI and ITBV on the day of enrollment to dif-ferentiate between ALI/ARDS patients and non-ALI/

ARDS patients. The area under the curve for PVPI(0.886; confidence interval, 0.836 to 0.935) was signifi-cantly larger than that for ITBV (0.575; confidenceinterval, 0.471 to 0.651; P < 0.01) (Figure 7).The cutoff value for the definitive quantitative diagno-

sis of ALI/ARDS needs to be determined consideringthe high specificity despite the decreased level of sensi-tivity, as discussed in the next section. The cutoff valueof the PVPI to diagnose ALI/ARDS was found to bebetween 2.85 (sensitivity, 0.54; specificity, 0.95) and 2.6(sensitivity, 0.64; specificity, 0.90). The cutoff value ofthe PVPI to diagnose non-ALI/ARDS was between 1.7(sensitivity, 0.50; specificity, 0.95) and 2.0 (sensitivity,0.70; specificity, 0.90).

Table 1 Patient characteristics

Variable All (n = 266) ALI/ARDS (n = 207) Cardiogenic edema (n = 26) Pleural effusion with atelectasis (n = 33)

Age (years) 67.3 ± 16.2 66.7 ± 16.8 70.0 ± 12.5 69.4 ± 14.3

Male 175 (65.3%) 134 (64.7%) 14 (53.8%) 27 (77.1%)

Height (cm) 159.8 ± 10.0 159.4 ± 10.0 157.8 ± 10.4 163.4 ± 9.3

Body weight (kg) 55.2 ± 10.6 54.8 ± 10.6 52.8 ± 11.1 59.2 ± 9.8

BSA (m2) 1.56 ± 0.20 1.55 ± 0.20 1.52 ± 0.21 1.64 ± 0.18

Body mass index 21.3 ± 1.8 21.3 ± 1.8 20.9 ± 1.9 22.0 ± 1.5

Heart rate (beats/minute) 102 ± 25 103 ± 22 95 ± 26 103 ± 37

MAP (mmHg) 77 ± 18 76 ± 17 80 ± 18 79 ± 20

Vasopressor 180 (70.7%) 148 (71.5%)* 10 (38.5%)## 22 (66.7%)

Sepsis 144 (54.1%) 128 (61.8%)*,# 4 (15.4%) 12 (36.3%)

APACHE II score (points) 22.7 ± 8.0 23.4 ± 8.1** 19.7 ± 6.5 20.7 ± 7.7

SOFA score (points) 10.4 ± 3.7 10.7 ± 3.6* 8.1 ± 3.9 10.3 ± 3.5

Respiratory 1.2 ± 1.2 1.2 ± 1.3 0.77 ± 1.1 1.5 ± 0.9

Coagulation 0.9 ± 1.2 0.9 ± 1.1 1.0 ± 1.5 0.9 ± 1.1

Liver 2.3 ± 1.7 2.5 ± 1.6* 1.3 ± 1.6## 2.4 ± 1.7

Cardiovascular 0.7 ± 1.0 0.7 ± 1.0 0.4 ± 0.8 0.8 ± 1.0

Central nervous system 3.0 ± 0.8 3.1 ± 0.8* 2.8 ± 0.7 2.7 ± 0.7

Renal 2.3 ± 1.4 2.4 ± 1.4 2.0 ± 1.3 2.0 ± 1.5

PEEP (cmH2O) 8.6 ± 4.5 8.7 ± 4.7 7.9 ± 4.2 8.9 ± 3.5

Plateau pressure (cmH2O) 22.1 ± 5.7 22.3 ± 5.6 20.6 ± 6.9 21.5 ± 5.2

PaO2/FIO2 ratio (mmHg) 155.3 ± 70.7 150.5 ± 70.9 166.7 ± 69.8 176.2 ± 67.5

Lung Injury Score (points) 2.3 ± 0.6 2.3 ± 0.6# 2.3 ± 0.6 1.8 ± 0.7

Ejection fraction (%) 55.1 ± 13.1* 56.4 ± 12.1* 46.8 ± 13.0 55.0 ± 17.0

Stroke volume variation (%) 15 ± 7 16 ± 7* 11 ± 6## 16 ± 7

Cardiac index (l/m2/minute) 3.4 ± 1.2 3.5 ± 1.3 3.0 ± 1.0 3.4 ± 1.2

ITBV (ml/m2) 1040 ± 303 1021 ± 257* 1312 ± 527## 948 ± 215

EVLWI (ml/kg) 16.8 ± 7.1 18.5 ± 6.8*,# 14.4 ± 4.0# 8.3 ± 2.1

PVPI 2.9 ± 1.4 3.2 ± 1.4*,# 2.0 ± 0.8 1.6 ± 0.5

Maximal EVLWI (ml/kg) 19.2 ± 8.5 21.2 ± 8.1*,# 16.4 ± 7.5# 9.4 ± 2.1

Maximal PVPI 3.4 ± 1.6 3.7 ± 1.6*,# 2.2 ± 1.2 1.9 ± 0.6

VFD within 28 days (days) 11.3 ± 10.3 9.8 ± 9.9* 20.3 ± 8.0## 13.1 ± 10.7

28-day mortality 99 (37.2%) 84 (40.6%) 7 (26.9%) (24.2%)

Data presented as mean ± standard deviation or n (%). Data on the day of enrollment were presented, except for maximal extravascular lung water index(EVLWI), and maximal pulmonary vascular permeability index (PVPI) during the 3-day study period. ALI/ARDS, acute lung injury/acute respiratory distresssyndrome; APACHE, Acute Physiology and Chronic Health Evaluation; BSA, body surface area; ITBV, intrathoracic blood volume index; MAP, mean arterialpressure; PEEP, positive end-expiratory pressure; SOFA, Sequential Organ Failure Assessment; VFD, ventilator-free days. *P < 0.01 vs. cardiogenic edema, **P <0.05 vs. cardiogenic edema, #P < 0.01 vs. pleural effusion with atelectasis, ##P < 0.05 vs. pleural effusion with atelectasis.


of 15

DiscussionIn this prospective multi-institutional observational study,EVLW and pulmonary vascular permeability were assessedby transpulmonary thermodilution in patients requiringmechanical ventilation with P/F ratio ≤ 300 mmHg andbilateral pulmonary infiltration on chest X-ray scan. Theresults showed that EVLW was greater in patients withALI/ARDS and cardiogenic edema than in those withpleural effusion with atelectasis; that pulmonary vascularpermeability was increased in patients with ALI/ARDScompared with cardiogenic edema and pleural effusionwith atelectasis patients; and that EVLW, the crucialpathogenetic factor of pulmonary edema, was weakly cor-related with the P/F ratio in patients with ALI/ARDS andcardiogenic edema.ARDS is associated with a high incidence of morbidity

and mortality despite the development of improvedmanagement techniques over the past two decades [45].Difficulties in selecting appropriate patient populationsthat would benefit from specific treatments occur largelybecause of the lack of homogeneity in the disease defini-tion. The AECC criteria, which have been exclusivelyused as the inclusion criteria in clinical trials for ALI/ARDS, were designed to identify patients with ALI/ARDS [9]. These criteria are inclusive, so that the popu-lation selected on the basis of these criteria can be veryheterogeneous in disease severity and clinical outcomes.This heterogeneity might be the reason why most clini-cal trials in ARDS have failed to achieve improved mor-tality. There is no validated biomarker that allows early

recognition of increased lung vascular permeability, thehallmark of ALI/ARDS pathogenesis. Although the Ber-lin definition is expected to have sufficient specificityand acceptable sensitivity, an accurate definition canalso help clinicians identify patients who may benefitfrom precise ventilatory strategies, diagnostic proce-dures, or drug therapies.Although pulmonary capillary hydrostatic pressure

elevation and pulmonary vascular permeability increaseare known to induce pulmonary edema, discrimination ofthese pathogeneses is important because of the differencein treatments. At present, the differential diagnosis ismade on the basis of the assessment of left atrial pressure,which is assumed to be normal in ALI/ARDS [46]. How-ever this hemodynamic definition of ALI/ARDS is contro-versial, as suggested previously [47]. The pulmonary arteryocclusion pressure might not reflect the hydrostatic pres-sure in the pulmonary micro-vessels [48,49] and cannot beaccurately measured [50,51]. Moreover, left ventricularpreload might be elevated in ALI/ARDS patients, espe-cially in those who have already received volume resuscita-tion and/or pre-existing or sepsis-related cardiacdysfunction, as described in the Berlin definition [10]. Thepulmonary artery occlusion pressure was recently found tobe elevated in 30% of ARDS patients [52]. The definitionof ALI/ARDS should thus include the functional featuresof the pathophysiology of this syndrome; that is, increasedpulmonary microvascular permeability [46]. The PVPI hasbeen suggested to be an important parameter in ALI/ARDS pathogenesis [53].

EVLWI in ALI/ARDS, cardiogenic edema, and pleuraleffusion with atelectasisA close agreement between thermodilution EVLWvalues and gravimetric lung water measurements, whichare thought to be the gold standard for the quantitativeevaluation of EVLW, has been shown in animal modelswith lung injury [27,54,55] and in a human autopsystudy [25].In this study, EVLWI ≥ 10 ml/kg was defined as pul-

monary edema. Although no definitive quantitative cri-teria of EVLWI for pulmonary edema were established,we recently reported that the normal EVLWI value isapproximately 7.4 ± 3.3 ml/kg from a human autopsystudy and this value can distinguish between healthyand pathological lungs [24]. Because EVLWI ≥ 10 ml/kghas been shown to predict progression to acute lunginjury, this value was used for defining pulmonaryedema in this study [1,29]. Only three patients hadEVLWI < 10 ml/kg from the 155 ARDS patients, basedon the AECC definition.Regarding the effect of pleural effusion on EVLW

measurements, Blomqvist and colleagues showed thatthe pleural fluid level did not affect the reliability of the

Table 2 Underlying conditions and mechanism forpatients with acute lung injury/acute respiratory distresssyndrome

Sepsis (n) 128

Respiratory 83

Abdomen 26

Urinary tract 5

Soft tissue 3

Others 11

Non-infectious condition (n) 79

Trauma 11

Surgery 8

Burn 7

Respiratory diseases 10

Gastrointestinal diseases 5

Central nervous system 4

Severe acute pancreatitis 3

Vascular diseases 2

Others 29

Direct or indirect lung injury

Direct injury 125

Indirect injury 82


of 15

double-indicator dilution technique for measuringEVLW in dogs [56]. This relationship has been alsodemonstrated using the single indicator thermodilutionmethod in critically ill patients in a previous study [57].We showed a very close correlation between premortemsingle-indicator thermodilution measurement of EVLWand postmortem lung weight, regardless of the degree ofpleural effusion [25]. In this study, the EVLWI ofpatients with pleural effusion with atelectasis was lessthan the defined value of pulmonary edema, despite thefact that patients had P/F ratio ≤ 300 mmHg and bilat-eral infiltration. This suggests that the EVLWI value

might not be significantly influenced by pleural effusionwith or without massive atelectasis. If patients haveEVLWI < 10 ml/kg but P/F ratio ≤ 300 mmHg andbilateral infiltration, thoracic ultrasound for the evalua-tion of pleural effusion should be performed.The EVLWI value was different between ALI/ARDS

and cardiogenic edema in this study. Hydrostatic pressureelevation occurred without a significant increase in vascu-lar permeability in patients with cardiogenic edema,whereas both increased vascular permeability and ele-vated hydrostatic pressure contributed to the develop-ment of edema in ALI/ARDS patients. As already

**

: p < 0.01 vs. pleural effusion with atelectasis** : p < 0.01 vs. cardiogenic edema

**

A

B

EVLW

I-0m

ax E

VLW

I

Figure 2 Comparison of extravascular lung water indexed to predicted body weight. Comparison of extravascular lung water indexed topredicted body weight of patients with acute lung injury (ALI)/acute respiratory distress syndrome (ARDS), cardiogenic edema, and pleuraleffusion with atelectasis on the day of enrollment and the maximal value during the study period. (A) Extravascular lung water indexed topredicted body weight (EVLWI) on the day of enrollment was significantly higher in ALI/ARDS patients than in pleural effusion with atelectasispatients and cardiogenic edema patients. EVLWI was also higher in cardiogenic edema patients than in pleural effusion with atelectasis patients.(B) Differences were found when the maximal EVLWI value was compared between day 0 and day 2. Data presented as median (interquartilerange). *P < 0.01 vs. pleural effusion with atelectasis. **P < 0.01 vs. cardiogenic edema. EVLWI-0, extravascular lung water index on day ofenrollment; maxEVLWI, maximal value of extravascular lung water index from days 0 to 2.


of 15

suggested, left ventricular preload can be elevated in ALI/ARDS, especially in patients who already have volumeresuscitation and/or have pre-existing or sepsis-relatedcardiac dysfunction [10,52]. In this regard, more than15% of the ALI/ARDS patients included in this study hada combined mechanism of an elevated left ventricularpreload and vascular permeability; this may have contrib-uted to the difference in EVLWI between the two groups.One should also consider that the heterogeneity of thepatients’ condition may affect the value of the EVLWI, asshown in a recent meta-analysis where the EVLWI inseptic patients was higher than in surgical patients(11.0 ml/kg vs. 7.2 ml/kg) [58].

Determination of the cutoff value for ALI/ARDS diagnosisAlthough EVLW has been extensively evaluated as notonly a predictor but also a diagnostic and prognosticparameter for ALI/ARDS [59], this value provides onlythe degree of accumulated EVLW and not the underly-ing pathophysiological mechanism.Increased pulmonary vascular permeability can be esti-

mated by the PVPI [53]. A recent retrospective studyshowed EVLWI values of 22 ± 9 and 16 ± 4 in patientswith ARDS and hydrostatic pulmonary edema, respectively[7]. On the contrary, in our study, these values were 18.5 ±6.8 and 8.3 ± 2.1, respectively. These differences mighthave influenced the PVPI values obtained in both studies

: p < 0.01 vs. pleural effusion with atelectasisand cardiogenic edema

A

B

Figure 3 Comparison of pulmonary vascular permeability index. Comparison of pulmonary vascular permeability index (PVPI) of patientswith acute lung injury (ALI)/acute respiratory distress syndrome (ARDS), cardiogenic edema, and pleural effusion with atelectasis on the day ofenrollment and the maximal value during the study period. (A) PVPI was higher in ALI/ARDS patients than in cardiogenic edema and pleuraleffusion with atelectasis patients. There was no difference in the index between cardiogenic edema and pleural effusion with atelectasis patients.(B) Differences were found when the maximal index value was compared between day 0 and day 2. Data presented as median (interquartilerange). *P < 0.01 vs. pleural effusion with atelectasis and cardiogenic edema. PVPI-0, pulmonary vascular permeability index on the day ofenrollment; maxPVPI, maximal value of pulmonary vascular permeability index from days 0 to 2.


of 15

A B

EVLW

I-0

EVLW

I-0ITBV-0

Figure 4 Extravascular lung water index and pulmonary vascular permeability index/intrathoracic blood volume correlation in ALI/ARDS patients. Correlation between extravascular lung water index (EVLWI) and pulmonary vascular permeability index (PVPI) and that betweenEVLWI and intrathoracic blood volume (ITBV) in patients with acute lung injury/acute respiratory distress syndrome. There was a strongcorrelation between EVLWI and PVPI (r = 0.729, P < 0.01) (A) and a weak correlation between EVLWI and ITBV (r = 0.236, p < 0.01) (B). EVLWI-0,extravascular lung water index on the day of enrollment; PVPI-0, pulmonary vascular permeability index on the day of enrollment; ITBV-0,intrathoracic blood volume on the day of enrollment.

A B

EVLW

I-0

EVLW

I-0

ITBV-0Figure 5 Extravascular lung water index and pulmonary vascular permeability index/intrathoracic blood volume correlation in non-ALI/ARDS patients. Correlation between extravascular lung water index (EVLWI) and pulmonary vascular permeability index (PVPI) and thatbetween EVLWI and intrathoracic blood volume (ITBV) in patients with non-acute lung injury (ALI)/acute respiratory distress syndrome (ARDS).For this analysis, cardiogenic edema and pleural effusion with atelectasis patients were combined as non-ALI/ARDS. EVLWI had a moderatecorrelation with PVPI (r = 0.464, P < 0.01) (A) and with ITBV (r = 0.493, P < 0.01) (B). EVLWI-0, extravascular lung water index on the day ofenrollment; PVPI-0, pulmonary vascular permeability index on the day of enrollment; ITBV-0, intrathoracic blood volume on the day ofenrollment.


of 15

(4.7 ± 1.8 vs. 3.2 ± 1.4 in ALI/ARDS and 2.1 ± 0.5 vs. 2.0 ±0.8 in cardiogenic edema). Although the previous studydid not show the Acute Physiology and Chronic HealthEvaluation II score and mortality, the difference in diseaseseverity and EVLWI values for inclusion or definition ofpulmonary edema might have influenced the determinedcutoff value for ALI/ARDS. However, the proposed cut-offvalue of PVPI ≥ 3 for the diagnosis of ALI/ARDS wasalmost the same as the present study, suggesting that thismulticenter study confirmed the proposed value.The cutoff value for ALI/ARDS should be determined by

considering the balance between sensitivity and specificity.Clinical trials focusing on specific therapy for ALI/ARDSshould select cutoff values with higher specificity, whereasthose focusing on the early recognition of this conditionand therapeutic intervention should select cutoff valueswith high sensitivity. Because the procedures for measur-ing EVLW and pulmonary vascular permeability are notnon-invasive, the validity of the definite diagnosis of ALI/ARDS made on the basis of PVPI values that have highspecificity should be considered.

LimitationsThe mechanism underlying respiratory insufficiency -that is, permeability pulmonary edema, cardiogenicedema, or pleural effusion with atelectasis, which maybe exclusive or overlapped - was defined by expert con-sensus, and subjectivity therefore cannot be completelyruled out. Nonetheless, only those patients who wereconsidered eligible by all the experts were included in

this study and 35 patients were excluded from the ana-lysis. There may have been some statistical bias in thisregard and the patient population may not represent thegeneral population of mechanically ventilated patientswith hypoxemic respiratory failure and radiographicinfiltrates. Fourteen of the 35 excluded patients werejudged to have respiratory failure secondary to sepsis-induced increased pulmonary vascular permeability buthad to be excluded because they presented EVLWI< 10 ml/kg. This exclusion may bias the study.Although the study sample size was not small, the num-

ber of patients with cardiogenic pulmonary edema wasfewer than those with ALI/ARDS. Most patients with car-diogenic pulmonary edema are managed using non-inva-sive positive pressure ventilation without trachealintubation; hence, these patients could not be included inthis study. There may have been some statistical bias inthis regard.Pulmonary inflammation (that is, pneumonia) needs to

be considered because it might influence the thermodi-lution technique findings. Inflamed cells and purulentmatter, including microabscesses, may increase lungweight despite increased EVLW. Further evaluation maybe required to assess ALI/ARDS secondary to directinjury from pneumonia.EVLWI ≥ 10 ml/kg was used for defining pulmonary

edema in this study. Although no definitive quantitativecriteria for pulmonary edema were established, thisvalue was selected on the basis of the value found inour recent human autopsy study and those used for

ALI/ARDS (n =207) Cardiogenic edema (n =26) All patients except for pleural effusion with atelectasis (n =233)

EVLWI-0 EVLWI-0 EVLWI-0

A B C

Figure 6 Correlation between extravascular lung water index and PaO2/FiO2 ratio. There was a negative and weak correlation betweenextravascular lung water index (EVLWI) and PaO2/FiO2 (P/F) ratio in all patients except for pleural effusion with atelectasis patients (r = -0.213, P< 0.01) (A). In acute lung injury (ALI)/acute respiratory distress syndrome (ARDS) patients (n = 207), a weak correlation between EVLWI and P/Fratio was observed (r = -0.215, P < 0.01) (B). No correlation between EVLWI and pulmonary vascular permeability index was observed incardiogenic edema patients (n = 26; r = -0.176, P = 0.39) (C). EVLWI-0, extravascular lung water index on the day of enrollment; P/F-0, PaO2/FiO2

ratio on the day of enrollment.


of 15

defining pulmonary edema in previously reported stu-dies. Lowering the EVLWI value for pulmonary edemamay have led to the inclusion of less severe pulmonaryedema cases and might have influenced the diagnosticcutoff value and its accuracy. However, most of thepatients who had pleural effusion with atelectasisshowed EVLWI < 10 ml/kg, suggesting that the valuewas not too low.

ConclusionThis study showed that EVLW was greater in patientswith ALI/ARDS and cardiogenic edema than patients withpleural effusion with atelectasis; that pulmonary vascularpermeability was increased in patients with ALI/ARDScompared with cardiogenic edema and pleural effusionwith atelectasis patients; and that the cutoff value of PVPI

for the quantitative diagnosis of ALI/ARDS was between2.6 and 2.85, with a specificity of 0.9 to 0.95, and thatPVPI < 1.7 ruled out an ALI/ARDS diagnosis (specificity,0.95).

Key messages• EVLW was greater in patients with ALI/ARDS andcardiogenic edema than patients with pleural effu-sion with atelectasis.• Pulmonary vascular permeability was increased inpatients with ALI/ARDS compared with cardiogenicedema and pleural effusion with atelectasis patients.• The cutoff value of the PVPI for the quantitativediagnosis of ALI/ARDS was between 2.6 and 2.85,with a specificity of 0.9 to 0.95, and PVPI < 1.7ruled out an ALI/ARDS diagnosis (specificity, 0.95).

IITBV-0

PVPI-0

Sen

siti

vity

1 - Specificity Figure 7 Receiver operating characteristic curves for pulmonary vascular permeability index and intrathoracic blood volume. Receiveroperating characteristic curves for pulmonary vascular permeability index (PVPI) and intrathoracic blood volume (ITBV) on the day of enrollmentfor the differential diagnosis of acute lung injury/acute respiratory distress syndrome. The area under the curve for PVPI (0.886; confidenceinterval, 0.836 to 0.935) was significantly larger than that for ITBI (0.425; confidence interval, 0.359 to 0.529; P < 0.01). PVPI-0, pulmonary vascularpermeability index on the day of enrollment; ITBV-0, intrathoracic blood volume on the day of enrollment.


of 15

AbbreviationsAECC: American-European Consensus Conference; ALI: acute lung injury;ARDS: acute respiratory distress syndrome; BNP: B-type natriuretic peptide;CO: cardiac output; EVLW: extravascular lung water; EVLWI: extravascular lungwater index; FiO2: fraction of inspired oxygen; GEDV: global end-diastolicvolume; ITBV: intrathoracic blood volume; PaO2: partial pressure of arterialoxygen; P/F ratio: PaO2/FiO2 ratio; PVPI: pulmonary vascular permeabilityindex.

AcknowledgementsThis prospective, observational, multi-institutional study was approved by theethics committee of each of the 23 institutions; Tohoku University GraduateSchool of Medicine, St. Marianna University School of Medicine, KansaiMedical University, Nara Medical University, Kurume University School ofMedicine, Fukuoka University, Nippon Medical School Hospital, NihonUniversity School of Medicine Itabashi Hospital, Tokyo Medical and DentalUniversity Hospital of Medicine, Juntendo University Nerima Hospital, JikeiUniversity School of Medicine, National Hospital Organization DisasterMedical Center, Saiseikai Yokohamashi Tobu Hospital, Social InsuranceChukyo Hospital, Kansai Medical University Takii Hospital, Osaka City GeneralHospital, Kobe City Medical Center General Hospital, Hiroshima City Hospital,Yamaguchi University Hospital, Nagasaki University Hospital, Nippon MedicalSchool Tama Nagayama Hospital, Nippon Medical School Chiba HokusouHospital, and Aizu Chuo Hospital. This work was supported in part by aGrant-in-Aid for Scientific Research (22592023) from the Ministry ofEducation, Science, Sports, and Culture of Japan.

Author details1Division of Emergency Medicine, Tohoku University Graduate School ofMedicine, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan. 2Departmentof Emergency and Critical Care Medicine, St. Marianna University School ofMedicine, 2-16-1 Sugao, Miyamae, Kawasaki, Kanagawa 216-8511, Japan.3Department of Emergency and Critical Care Medicine, Kansai MedicalUniversity, 10-15 Fumizono-cho, Moriguchi City, Osaka 570-8506, Japan.4Department of Emergency and Critical Care Medicine, Nara MedicalUniversity, 840 Shinjo-cho, Kashihara, Nara 634-8521, Japan. 5Department ofEmergency and Critical Care Medicine, Kurume University School ofMedicine, 67 Asahi-machi, Kurume-shi, Fukuoka 830-0011, Japan.6Department of Emergency and Critical Care Medicine, Faculty of Medicine,Fukuoka University, 7-45-1 Nanakuma, Jonan-ku, Fukuoka City, Fukuoka 814-0180, Japan. 7Advanced Emergency and Critical Care Center, TohokuUniversity Hospital, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan.8Department of Emergency and Critical Care Medicine, Nippon MedicalSchool Hospital, 1-1-5 Sendagi, Bunkyo-ku, Tokyo 113-8603, Japan.9Department of Emergency and Critical Care Medicine, Aidu Chuo Hospital,1-1 Tsuruga, Aiduwakamatsu, Fukushima 965-8611, Japan. 10Division ofEmergency and Critical Care Medicine, Department of Acute Medicine,Nihon University School of Medicine, 30-1 Oyaguchi kamimachi, Itabashi-ku,Tokyo 173-8610, Japan. 11Shock Trauma and Emergency Medical Center,Tokyo Medical and Dental University Hospital, 1-5-45 Yushima, Bunkyo-ku,Tokyo 113-8519, Japan. 12Department of Emergency and Critical CareMedicine, Juntendo University Nerima Hospital, 3-1-10 Takanodai, Nerima-ku,Tokyo 177-8521, Japan. 13Department of Critical Care Medicine, JikeiUniversity School of Medicine, 3-19-18 Nishi-shinbashi, Minato-ku, Tokyo 105-8471, Japan. 14Emergency and Critical Care Medicine, National HospitalOrganization Disaster Medical Center, 3256 Midori-cho, Tachikawa-shi, Tokyo190-0014, Japan. 15Department of Intensive Care Medicine, SaiseikaiYokohamashi Tobu Hospital, 3-6-1 Shimosumiyosi, Tsurumi-ku, YokohamaCity, Kanagawa 230-8765, Japan. 16Department of Emergency and CriticalCare Medicine, Social Insurance Chukyo Hospital, 1-1-10 Sanjo, Minami-ku,Nagoya City, Aichi 457-8510, Japan. 17Advanced Emergency and Critical CareCenter, Kansai Medical University Takii Hospital, 10-15 Fumizono-machi,Moriguchi Cty, Osaka 570-8507, Japan. 18Emergency and Critical CareMedical Center, Osaka City General Hospital, 2-13-22, Miyakojima Hondori,Miyakojima, Osaka 534-0021, Japan. 19Department of Anesthesia, Kobe CityMedical Center General Hospital, 2-2-1 Minatojimaminamimachi, Chuo-ku,Kobe City, Hyogo 650-0046, Japan. 20Department of Anesthesia andIntensive Care, Hiroshima City Hospital, 7-33 Motomachi, Naka-ku, Hiroshima-shi, Hiroshima 730-8518, Japan. 21Advanced Medical Emergency and CriticalCare Center, Yamaguchi University Hospital, 1-1-1 Minamikogushi, Ube City,Yamaguchi 755-8505, Japan. 22Intensive Care Unit, Nagasaki University

Hospital, 1-7-1 Sakamoto, Nagasaki 852-8501, Japan. 23Department ofEmergency and Critical Care Medicine, Nippon Medical School TamaNagayama Hospital, 1-7-1 Nagayama, Tama-shi, Tokyo 206-8512, Japan.24Department of Emergency and Critical Care Medicine, Nippon MedicalSchool Chiba Hokusoh Hospital, 1715 Kamagari, Inzai-shi, Chiba 270-1694,Japan.

Authors’ contributionsAll authors conceived and designed the study, wrote the study protocol andcontributed to clinical data acquisition. The statistical analyses and the firstdraft of manuscript were performed by SK. All authors amended andcommented on the manuscript and approved the final version.

Competing interestsYT is a member of the medical advisory board of Pulsion Medical Systems.The remaining authors declare that they have no competing interests.

Received: 26 May 2012 Revised: 15 September 2012Accepted: 6 December 2012 Published: 11 December 2012

References1. Michard F: Bedside assessment of extravascular lung water by dilution

methods: temptations and pitfalls. Crit Care Med 2007, 35:1186-1192.2. Ware LB, Matthay MA: Acute pulmonary edema. N Engl J Med 2005,

353:2788-2796.3. Atabai K, Matthay MA: The pulmonary physician in critical care. 5: acute

lung injury and the acute respiratory distress syndrome: definitions andepidemiology. Thorax 2002, 57:452-458.

4. Rubenfeld GD, Caldwell E, Granton J, Hudson LD, Matthay MA:Interobserver variability in applying a radiographic definition for ARDS.Chest 1999, 116:1347-1353.

5. Meade MO, Cook RJ, Guyatt GH, Groll R, Kachura JR, Bedard M, Cook DJ,Slutsky AS, Stewart TE: Interobserver variation in interpreting chestradiographs for the diagnosis of acute respiratory distress syndrome. AmJ Respir Crit Care Med 2000, 161:85-90.

6. Lichtenstein D, Goldstein I, Mourgeon E, Cluzel P, Grenier P, Rouby JJ:Comparative diagnostic performances of auscultation, chestradiography, and lung ultrasonography in acute respiratory distresssyndrome. Anesthesiology 2004, 100:9-15.

7. Monnet X, Anguel N, Osman D, Hamzaoui O, Richard C, Teboul JL:Assessing pulmonary permeability by transpulmonary thermodilutionallows differentiation of hydrostatic pulmonary edema from ALI/ARDS.Intensive Care Med 2007, 33:448-453.

8. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M,LeGall JR, Morris A, Spragg R: Report of the American-Europeanconsensus conference on ARDS: definitions, mechanisms, relevantoutcomes and clinical trial coordination. The Consensus Committee.Intensive Care Med 1994, 20:225-232.

9. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M,Legall JR, Morris A, Spragg R: Report of the American-European consensusconference on ARDS: definitions, mechanisms, relevant outcomes andclinical trial coordination. The Consensus Committee. Am J Respir CritCare Med 1994, 149:818-824.

10. ARDS Definition Task Force, Ranieri VM, Rubenfeld GD, Thompson BT,Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS: Acute respiratorydistress syndrome: the Berlin Definition. JAMA 2012, 307:2526-2533.

11. Ferguson ND, Davis AM, Slutsky AS, Stewart TE: Development of a clinicaldefinition for acute respiratory distress syndrome using the Delphitechnique. J Crit Care 2005, 20:147-154.

12. Neff MJ: The epidemiology and definition of the acute respiratorydistress syndrome. Respir Care Clin N Am 2003, 9:273-282.

13. Estenssoro E, Dubin A, Laffaire E, Canales HS, Sáenz G, Moseinco M, Bachetti P:Impact of positive end-expiratory pressure on the definition of acuterespiratory distress syndrome. Intensive Care Med 2003, 29:1936-1942.

14. Phua J, Stewart TE, Ferguson ND: Acute respiratory distress syndrome 40years later: time to revisit its definition. Crit Care Med 2008, 36:2912-2921.

15. Martin GS, Eaton S, Mealer M, Moss M: Extravascular lung water in patientswith severe sepsis: a prospective cohort study. Crit Care 2005, 9:R74-R82.

16. Phillips CR, Smith SM: Predicted body weight-indexed extravascular lungwater is elevated in acute respiratory distress syndrome. Crit Care Med2009, 37:377-378.


of 15

http://www.ncbi.nlm.nih.gov/pubmed/17334247?dopt=Abstract



































17. Davey-Quinn A, Gedney JA, Whiteley SM, Bellamy MC: Extravascular lungwater and acute respiratory distress syndrome - oxygenation andoutcome. Anaesth Intensive Care 1999, 27:357-362.

18. Michard F, Fernandez-Mondejar E, Kirov MY, Malbrain M, Tagami T: A newand simple definition for acute lung injury. Crit Care Med 2012,40:1004-1006.

19. Halperin BD, Feeley TW, Mihm FG, Chiles C, Guthaner DF, Blank NE:Evaluation of the portable chest roentgenogram for quantitatingextravascular lung water in critically ill adults. Chest 1985, 88:649-652.

20. Lindqvist B: Experimental uraemic pulmonary oedema including: criteriafor pulmonary oedema in anuric rabbits, the role of uramia andoverhydration, and a literary survey on the problems of uraemicpulmonary oedema (fluid-retention lung, etc.). Acta Med Scand 1964,176(Suppl 418):1.

21. Mihm FG, Feeley TW, Rosenthal MH, Lewis F: Measurement ofextravascular lung water in dogs using the thermal-green dye indicatordilution method. Anesthesiology 1982, 57:116-122.

22. Mihm FG, Feeley TW, Jamieson SW: Thermal dye double indicator dilutionmeasurement of lung water in man: comparison with gravimetricmeasurements. Thorax 1987, 42:72-76.

23. Sakka SG, Rühl CC, Pfeiffer UJ, Beale R, McLuckie A, Reinhart K, Meier-Hellmann A: Assessment of cardiac preload and extravascular lung waterby single transpulmonary thermodilution. Intensive Care Med 2000,26:180-187.

24. Neumann P: Extravascular lung water and intrathoracic blood volume:double versus single indicator dilution technique. Intensive Care Med1999, 25:216-219.

25. Tagami T, Kushimoto S, Yamamoto Y, Atsumi T, Tosa R, Matsuda K,Oyama R, Kawaguchi T, Masuno T, Hirama H, Yokota H: Validation ofextravascular lung water measurement by single transpulmonarythermodilution: human autopsy study. Crit Care 2010, 14:R162.

26. Cordemans C, De laet I, Van Regenmortel N, Schoonheydt K, Dits H,Huber W, Malbrain ML: Fluid management in critically ill patients: therole of extravascular lung water, abdominal hypertension, capillary leakand fluid balance. Annals Intensive Care 2012, 2(Suppl 1):S1.

27. Katzenelson R, Perel A, Berkenstadt H, Preisman S, Kogan S, Sternik L,Segal E: Accuracy of transpulmonary thermodilution versus gravimetricmeasurement of extravascular lung water. Crit Care Med 2004,32:1550-1554.

28. Schuster DP: The search for ‘objective’ criteria of ARDS. Intensive Care Med2007, 33:400-402.

29. LeTourneau JL, Pinney J, Phillips CR: Extravascular lung water predictsprogression to acute lung injury in patients with increased risk. Crit CareMed 2012, 40:847-854.

30. Monnet X, Persichini R, Ktari M, Jozwiak M, Richard C, Teboul JL: Precisionof the transpulmonary thermodilution measurements. Crit Care 2011, 15:R204.

31. Tagami T, Kushimoto S, Tosa R, Omura M, Hagiwara J, Hirama H, Yokota H:The precision of PiCCO® measurements in hypothermic post-cardiacarrest patients. Anesthesia 2012, 67:236-243.

32. Newman EV, Merrell M, Genecin A, Monge C, Milnor WR, McKeever WP: Thedye dilution method for describing the central circulation. An analysis offactors shaping the time-concentration curves. Circulation 1951,4:735-746.

33. Michard F, Schachtrupp A, Toens C: Factors influencing the estimation ofextravascular lung water by transpulmonary thermodilution in criticallyill patients. Crit Care Med 2005, 33:1243-1247.

34. In Handbook of Clinical Drug Data.. 7 edition. Edited by: Knoben JE,Anderson PO. Hamilton, IL: Drug Intelligence; 1993:.

35. Crapo RO, Morris AH, Gardner RM: Reference spirometric values usingtechniques and equipment that meet ATS recommendations. Am RevRespir Dis 1981, 123:659-664.

36. O’Brien JM Jr, Phillips GS, Ali NA, Lucarelli M, Marsh CB, Lemeshow S: Bodymass index is independently associated with hospital mortality inmechanically ventilated adults with acute lung injury. Crit Care Med 2006,34:738-744.

37. Berkowitz D, Danai P, Eaton S, Moss M, Martin GS: Accuratecharacterization of extravascular lung water in acute respiratory distresssyndrome. Crit Care Med 2008, 36:1803-1809.

38. Acute Respiratory Distress Syndrome Network: Ventilation with lower tidalvolumes as compared with traditional tidal volumes for acute lung

injury and the acute respiratory distress syndrome. N Engl J Med 2000,342:1301-1308.

39. Phillips CR, Chesnutt MS, Smith SM: Extravascular lung water in sepsis-associated acute respiratory distress syndrome: indexing with predictedbody weight improves correlation with severity of illness and survival.Crit Care Med 2008, 36:69-73.

40. Craig TR, Duffy MJ, Shyamsundar M, McDowell C, McLaughlin B, Elborn JS,McAuley DF: Extravascular lung water indexed to predicted body weightis a novel predictor of intensive care unit mortality in patients withacute lung injury. Crit Care Med 2010, 38:114-120.

41. Knaus WA, Draper EA, Wanger DP, Zimmerman JE: APACHE II: a severityclassification system. Crit Care Med 1985, 13:818-829.

42. Ferreira FL, Bota DP, Bross A, Mélot C, Vincent JL: Serial evaluation of theSOFA score to predict outcome in critically ill patients. JAMA 2001,286:1754-1758.

43. Murray JF, Matthay MA, Luce JM, Flick MR: An expanded definition of theadult respiratory distress syndrome. Am Rev Respir Dis 1998, 138:720-723.

44. Bakhtiari K, Meijers JCM, de Jonge E, Levi M: Prospective validation of theInternational Society of Thrombosis and Haemostasis scoring system fordisseminated intravascular coagulation. Crit Care Med 2004, 32:2416-2421.

45. Rubenfeld GD, Caldwell E, Peabody E, Weaver J, Martin DP, Neff M, Stern EJ,Hudson LD: Incidence and outcomes of acute lung injury. N Engl J Med2005, 353:1685-1693.

46. Abraham E, Matthay MA, Dinarello CA, Vincent JL, Cohen J, Opal SM,Glauser M, Parsons P, Fisher CJ Jr, Repine JE: Consensus conferencedefinitions for sepsis, septic shock, acute lung injury, and acuterespiratory distress syndrome: time for a reevaluation. Crit Care Med2000, 28:232-235.

47. Isakow W, Schuster DP: Extravascular lung water measurements andhemodynamic monitoring in the critically ill: bed-side alternatives to thepulmonary artery catheter. Am J Physiol Lung Cell Mol Physiol 2006, 291:L1118-L1131.

48. Iberti TJ, Daily EK, Leibowitz AB, Schecter CB, Fischer EP, Silverstein JH:Assessment of critical care nurses’ knowledge of the pulmonary arterycatheter. The Pulmonary Artery Catheter Study Group. Crit Care Med1994, 22:1674-1678.

49. Nunes S, Ruokonen E, Takala J: Pulmonary capillary pressures during theacute respiratory distress syndrome. Intensive Care Med 2003,29:2174-2179.

50. Monnet X, Teboul JL: Invasive measures of left ventricular preload. CurrOpin Crit Care 2006, 12:235-240.

51. Iberti TJ, Fischer EP, Leibowitz AB, Panacek EA, Silverstein JH, Albertson TE:A multicenter study of physicians’ knowledge of the pulmonary arterycatheter. Pulmonary Artery Catheter Study Group. JAMA 1990,264:2928-2932.

52. National Heart, Lung, and Blood Institute Acute Respiratory DistressSyndrome (ARDS) Clinical Trials Network, Wiedemann HP, Wheeler AP,Bernard GR, Thompson BT, Hayden D, deBoisblanc B, Connors AF Jr,Hite RD, Harabin AL: Comparison of two fluid-management strategies inacute lung injury. N Engl J Med 2006, 354:2564-2575.

53. Schuster DP: The search for ‘objective’ criteria of ARDS. Intensive Care Med2007, 33:400-402.

54. Kirov MY, Kuzkov VV, Kuklin VN, Waerhaug K, Bjertnaes LJ: Extravascularlung water assessed by transpulmonary single thermodilution andpostmortem gravimetry in sheep. Crit Care 2004, 8:R451-R458.

55. Fernández-Mondéjar E, Castaño-Pérez J, Rivera-Fernández R, Colmenero-Ruiz M, Manzano F, Pérez-Villares J, de la Chica R: Quantification of lungwater by transpulmonary thermodilution in normal and edematouslung. J Crit Care 2003, 18:253-258.

56. Blomqvist H, Wickerts CJ, Rosblad PG: Effects of pleural fluid and positiveend-expiratory pressure on the measurement of extravascular lungwater by the double-indicator dilution technique. Acta Anaesthesiol Scand1991, 35:578-583.

57. Deeren DH, Dits H, Daelemans R, Malbrain ML: Effect of pleural fluid onthe measurement of extravascular lung water by single transpulmonarythermodilution. Clin Intensive Care 2004, 15:119-122.

58. Eichhorn V, Goepfert MS, Eulenburg C, Malbrain ML, Reuter DA:Comparison of values in critically ill patients for global end-diastolicvolume and extravascular lung water measured bytranscardiopulmonary thermodilution: a metaanalysis of the literature.Med Intensiva 2012.


of 15
























































































59. van der Heijden M, Groeneveld VBJ: Extravascular lung water to bloodvolume ratios as measuresof pulmonary capillary permeability innonseptic critically ill patients. J Crit Care 2010, 25:16-22.

doi:10.1186/cc11898Cite this article as: Kushimoto et al.: The clinical usefulness ofextravascular lung water and pulmonary vascular permeability index todiagnose and characterize pulmonary edema: a prospective multicenterstudy on the quantitative differential diagnostic definition for acutelung injury/acute respiratory distress syndrome. Critical Care 201216:R232.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit


of 15




RESEARCH Open Access The clinical usefulness of ...

Documents

Transcript of RESEARCH Open Access The clinical usefulness of ...